The neo4j-agent-memory Technology Compatibility Kit (TCK) provides a formal specification, executable test suite, and compliance framework for the neo4j-agent-memory ecosystem.
This documentation follows the Diataxis framework, organized into four sections:
| Section | Purpose |
|---|---|
Step-by-step lessons for getting started |
|
Goal-oriented recipes for specific tasks |
|
Technical descriptions of the machinery |
|
Background and context for understanding the system |
Tutorials
Learning-oriented guides that walk you through a complete task from start to finish.
-
Getting Started with the TCK — Install the TCK, write your first adapter, and run the test suite
-
Building Your First TypeScript Agent with Memory — Use
@neo4j-labs/agent-memoryto add memory to a Vercel AI SDK application
How-To Guides
Task-oriented guides for accomplishing specific goals.
-
Writing a TCK Adapter — Implement
BaseAdapterfor your memory system -
Cross-Language Conformance Testing — Test TypeScript, Go, C#, or R implementations via the HTTP bridge
-
Adding TCK to Your CI Pipeline — Run conformance tests in GitHub Actions
-
Obtaining TCK Certification — Generate compliance reports and submit to the registry
Reference
Technical descriptions of types, protocols, and interfaces.
-
BaseAdapter Interface Reference — All 26 abstract methods with signatures, parameters, and return types
-
HTTP Bridge Protocol Specification — Endpoints, request/response schemas, and data types
-
Compliance Tier Definitions — Bronze, Silver, and Gold requirements
-
Scenario ID Registry Format — Stable scenario ID scheme and registry structure
Explanation
Background concepts and architectural decisions.
-
TCK Architecture and Design Decisions — Why pytest over Gherkin, the adapter pattern, and the bridge protocol
-
The Agent Memory Model — Short-term, long-term, and reasoning memory explained
-
Cross-Agent Memory Sharing — How agents in different languages share the same Neo4j graph